sign in sign up
<<
Home , SN 2014J, Type Ia supernova

Remnants as Clues to Their Progenitors

Daniel Patnaude and Carles Badenes

Abstract Supernovae shape the , chemically enrich their host , and generate powerful interstellar shocks that drive future generations of formation. The shock produced by a supernova event acts as a type of time ma- chine, probing the mass loss history of the progenitor system back to ages of ∼ 10 000 years before the explosion, whereas supernova remnants probe a much earlier stage of , interacting with material expelled during the progenitor’s much earlier evolution. In this chapter we will review how observations of super- remnants allow us to infer fundamental properties of the progenitor system. We will provide detailed examples of how bulk characteristics of a remnant, such as its chemical composition and dynamics, allow us to infer properties of the progeni- tor evolution. In the latter half of this chapter, we will show how this exercise may be extended from individual objects to SNR as classes of objects, and how there are clear bifurcations in the dynamics and spectral characteristics of core collapse and thermonuclear supernova remnants. We will finish the chapter by touching on re- cent advances in the modeling of massive , and the implications for observable properties of supernovae and their remnants.

1 Introduction

Supernova remnants provide a unique view of the supernova (SN) phenomenon. The hundreds of extragalactic supernovae (SNe) discovered each year by modern

Daniel Patnaude Smithsonian Astrophysical Observatory, 60 Garden St, Cambridge, MA 02138, e-mail:

arXiv:1702.03228v1 [astro-ph.HE] 10 Feb 2017 [email protected] Carles Badenes Department of Physics and Astronomy, and Pittsburgh Particle Physics, Astrophysics, and Cosmol- ogy Center (PITT-PACC), University of Pittsburgh, 3931 O’Hara St, Pittsburgh, PA 15260 e-mail: [email protected]

1 2 Daniel Patnaude and Carles Badenes surveys are unresolved point sources which fade before they can interact with any circumstellar material outside of their most immediate surroundings. Before they fade, they probe the evolution of the progenitor on timescales of ∼ 100–1000 years before the explosion. This is because the typical outflow speeds from the progenitor are ∼ 10–1000 km s−1, 100-1000× slower than the typical supernova blastwave velocity. In contrast, supernova remnants (SNRs) in the Local Group galaxies can be re- solved in some detail as they interact with the bulk of the circumstellar material expelled by their progenitors during pre-SN mass loss over much longer timescales. Additionally, SNRs are surrounded by resolved stellar populations that spawned their SN progenitors (Badenes et al., 2009; Williams et al., 2014; Jennings et al., 2014), and they have become prime hunting grounds in the search for their surviv- ing binary companions (e.g., Edwards et al., 2012; Kerzendorf et al., 2014). X-ray spectra of SNRs can reveal emission from neutron-rich isotopes in the ejecta which are sensitive probes of SN explosion mechanisms (e.g., Yamaguchi et al., 2015). In short, SNR observations can constrain aspects of SNe and their progenitors that are either difficult or impossible to access by optical SN observations. In this chapter, we review the clues that supernova remnants provide about their progenitor’s evolution and ultimate fate. We begin by highlighting studies of both Type Ia and core-collapse SNRs which inferred properties of their progenitors and evolution. We then discuss recent techniques that use bulk SNR observables to con- strain SN progenitor properties. We conclude by touching on recent results in mas- sive star evolution which may be relevant towards understanding observable prop- erties in supernova remnants. For further background and a different perspective of these topics, see the reviews by Badenes (2010) and Vink (2012).

2 - Progenitor Connections

The main challenge associated with studying supernova remnants is related to their type. Classically, SNe have been typed by their optical spectrum around maximum light (Filippenko, 1997). In general terms, Type Ia supernovae show no hydrogen in their optical spectra, with Type Ia (thermonuclear) SNe exhibiting strong Si II absorption around 615 nm. Type Ib/c are core-collapse supernovae, and differ from Ia SNe in that they possess either a weak or non-existent absorption feature, and little (Type Ib) or no (Ic) absorption from He I. The amount of helium in the optical spectrum may provide insight into the degree of mass-loss in the progenitor prior to core-collapse. In contrast, Type II SNe do show hydrogen in their spectra, and are further subdivided into types IIP, IIL, IIb and IIn, based on the evolution and characteristics of their or, in the case of Type IIn SNe, their optical spectrum. While SNe are common and somewhat accessible to typing, SNRs are evolved objects, and a priori it is non-trivial to associate them with each of the established SN types, though the detection of optical light echoes from the SN event do provide Supernova Remnants as Clues to Their Progenitors 3 clues in some cases (Rest et al., 2008). However, given the wealth of data available, a challenging opportunity exists to connect remnant properties back to the properties of the progenitor. Connecting remnant properties to those of the progenitor often involves a detailed analysis of the remnant’s broadband spectrum, combined with comparisons to hy- drodynamical models for supernova remnant ejecta evolution. In some cases such as those discussed below, this has produced interesting constraints. However, bulk properties such as emission line strengths, energy centroids, and chemical abun- dances depend on different aspects of a SN explosion than those of optical SNe. These properties can act as crucial benchmarks for specific stellar evolution sce- narios. In § 3 of this chapter, we explore how these so called “bulk observables” connect directly to the evolution of the progenitor, and how theorists and observers are connecting detailed stellar evolution models to properties of young supernova remnants. When studying SNRs, both as individual objects and as a class, the main proper- ties to consider are the chemical composition of the ejecta and the dynamics of the blast wave. The spectrum of a SNR in the X-ray band is a combination of nonthermal emission from the blast wave, and thermal emission from both swept-up circumstel- lar material and shocked supernova ejecta. The chemical composition of the shocked ejecta is determined by fitting models to the observed X-ray spectrum, which often requires some knowledge of the SNR dynamics (Borkowski et al., 2001). In turn, the blast wave dynamics are inferred by relying on critical assumptions about the re- lationship between the shock speed and the measured electron temperature. Only in certain cases, such as in the Galactic SNR Cas A, are the dynamics directly observed (DeLaney & Rudnick, 2003; Patnaude & Fesen, 2009). In most cases, the morphology is compared to self-similar models for blast wave evolution, which again require assumptions about the remant’s age, energetics, and circumstellar environment. These methods are vulnerable to circular logic and can result in detailed statements about the properties of a given SNR that fail to properly account for the considerable uncertainties introduced by the models. In other words, the critical assumptions made in connecting spectral properties to dynamics often compromise the the connections made back to the progenitor evolution. As an example, specific stellar evolution scenarios from SN Ia progenitors pre- dict circumstellar structures that range from the ρ ∝ r−2 profiles that are the sig- of isotropic mass-loss to relatively large low-density energy-driven cavities excavated by fast outflows from accreting white dwarfs (Badenes et al., 2007; Shen et al., 2013). As seen in Figure 1, the prompt radio and X-ray emission from SNe (Chomiuk et al., 2016; Margutti et al., 2014), weeks or months after the explosion, can only probe circumstellar material (CSM) structures up to a few hundred AU, composed of material expelled by the progenitor hundreds to thousands of years before the explosion, depending on the outflow velocity. In contrast, SNRs that are hundreds or thousands of years old can probe CSM structures up to several pc in size, expelled by the progenitor several thousands to millions of years before the explosion. These spatial scales are representative of typical stellar wind blown bubbles (Koo & McKee, 1992), and the associated 4 Daniel Patnaude and Carles Badenes temporal scales are directly relevant to stellar evolution scenarios for Type Ia SN progenitors (Wang & Han, 2012). For CC SNRs, the argument is even stronger, as mass-loss is often the single most uncertain ingredient in stellar evolution models for SN progenitors (Smith, 2014).

Fig. 1 Schematic circumstellar density profiles for isotropic SN Ia progenitor outflows (colored −1 3 −1 plots, dashed for uw = 20 km s ; solid for uw = 10 km s ), compared to the typical densities in the warm phase of the interstellar medium (horizontal grey band) (Ferriere,` 2001). The rulers on top of the plots show the time before the explosion (in yr) when the material at that radius was ejected by the SN progenitor. The rulers on the bottom show the typical sizes of circumstellar cavities excavated by fast outflows (Badenes et al., 2007; Shen et al., 2013). The radii probed by X-ray/radio follow-up observations of SNe and studies of Type Ia SNRs are shown by blue and red shaded vertical stripes. The blue arrows and red circles show upper limits for two Type Ia SNe (Chomiuk et al., 2016; Perez-Torres´ et al., 2014) and measurements for three Type Ia SNRs (Raymond et al., 2007; Slane et al., 2014; Borkowski et al., 2014).

A different approach is instead of studying the minute details of individual ob- jects, one probes the general characteristics of SNRs to see how they relate to the expected characteristics of a model derived from a potential progenitor scenario. This approach has been successful in Type Ia SNRs, and is now beginning to be explored in core collapse scenarios.

2.1 Type Ia SNR – Progenitor Connections

Despite decades of efforts, the progenitors of Type Ia SNe remain unidentified. The absence of hydrogen and helium in the optical spectra of SN Ia, their asso- ciation with both young and old stellar populations, the absence of a detectable shock breakout in the early light curves, and the chemical composition of the ejecta, as inferred from optical and X-ray observations, imply that the exploding star is a Supernova Remnants as Clues to Their Progenitors 5 - (C/O) (WD). The exact events leading to the explosion are unclear, but after the is somehow ignited, a thermonuclear 56 runaway consumes most of the star, synthesizing 0.5-1.0M of Ni, which powers the optical lightcurve through the 56Ni → 56Co → 56Fe. Different SN Ia progenitor scenarios attempt to provide a physical explanation for the destabilization of the WD and the subsequent explosion. In the single-degenerate (SD) scenario, the WD grows slowly in mass by from a non-degenerate companion, until it gets close to the (MCh ≈ 1.4M ). In the double degenerate (DD) scenario, the WD explodes after a dynamical merger or a collision with another WD, forming a short-lived object whose mass does not necessarily have to be MCh. See Hillebrandt et al. (2013) and Maoz et al. (2014) for recent reviews on SN Ia progenitors. The so-called ‘SN Ia progenitor problem’ persists because even though there is a great deal of indirect evidence supporting both scenarios, a definite ‘smoking gun’ remains elusive (see Maoz et al., 2014, for a discussion). In the SD scenario, the WD −7 −1 accretes mass at a rate of M˙ = ∼10 M yr (Shen & Bildsten, 2007). At these rates, stable hydrogen burning can occur on the surface of the WD, which allows steady mass growth to MCh. Accretion at higher rates can lead to the expansion of the accretor and engulfment of the companion into a phase, which effectively halts the growth of the WD. Solutions to this problem have been proposed, however, in the form of fast optically thick accretion winds that regulate the accretion (Hachisu et al., 1996). Lower accretion rates can result in the accu- mulation of degenerate material on the WD, which will eventually ignite as a nova where much of the accreted material and potentially some of the accretor mass is ejected, effectively eroding the mass of the WD. In the DD scenario, a more massive white dwarf merges or collides with a lower- mass WD and this rapid episode of mass accretion can effectively avoid problems related to inefficient mass growth in the SD scenario, and result in a SN Ia ex- plosion, not necessarily close to MCh (van Kerkwijk et al., 2010). However, off- center ignition in the primary WD can result in the formation of a O/Ne WD and an accretion-induced collapse to a without a SN Ia explosion (Saio & Nomoto, 1985). Thus two key observational probes of the SN Ia progenitor problem are the total mass of the ejecta (i.e., whether it is close to MCh or not) and the presence or absence of a fast accretion wind outflow before the explosion. SNR studies can effectively constrain both ejecta mass and pre-explosion outflows. When a WD explodes close to MCh in the SD scenario, it will have a dense 8 −3 (ρ & 2×10 g cm ) core where efficient electron captures can take place. This results in a significant production of neutron-rich isotopes such as 58Ni and 55Mn in the innermost ∼0.2 M of the WD (Brachwitz et al., 2000). In contrast, a SN Ia explosion with a DD progenitor will not present this neutronized core, unless it happened exactly at MCh. X-ray observations of Type Ia SNR can reveal the presence or absence of this neutronized core by comparing diagnostic mass ratios like Ni/Fe and Mn/Fe to the predictions of both MCh and sub-MCh models (Park et al., 2013). In the evolved (a) 3C 397 0.03

Tycho 5.4 Z

0.01 Kepler Mn/Fe Z 6 Daniel Patnaude1.8 and Carles Badenes

MCh DDT Sub-MCh -3 ρDDT [g cm ] MWD [M ] 0.72 Z Galactic SNR 3C 397, Yamaguchi et al. (2015) found high ratios of Ni/Fe3.9 x 10 7 and Mn/Fe1.15 in the X-ray spectrum. As seen in Figure 2, the Ni/Fe and Mn/Fe ratios2.6 x 10 7 are among1.06 1.3 x 10 7 0.97 the highest reported for Type Ia SNRs, and strongly0.001 0.003 indicate that the progenitor must 1.0 x 10 7 0.88 0.18 Z have exploded close to MCh, which is more naturally explained by the SD scenario. 0.010.03 0.1 0.3 Ni/Fe

(a) 3C 397

0.03 (b) 3C 397

Tycho Z 5.4 Z 5.4 Kepler 0.01 Kepler Tycho 0.01 0.03

1.8 Z Mn/Fe Mn/Fe

1.8 Z 0.72 Z MCh DDT Sub-MCh -3 ρDDT [g cm ] MWD [M ] 0.72 Z 3.9 x 10 7 1.15 Both MCh DDT models 2.6 x 10 7 1.06 Blue: n-NSE region 7 1.3 x 10 0.97 Z 0.001 0.003 0.18 Red: NSE + Si burning 1.0 x 10 7 0.88 0.18 Z

0.010.03 0.1 0.3 0.001 0.003 0.010.03 0.1 0.3 Ni/Fe Ni/Fe

Fig. 2 (a) Ni/Fe and Mn/Fe mass ratios in SN Ia models, compared with the observed values

(b) 3C 397 in 3C 397 (magenta diamond and gray regions). Blue symbols represent MCh delayed- models with different values of ρDDT as indicated in the legend. Red symbols represent sub-MCh detonation models with WD massesZ given in the legend. Models with the same progenitor metal- Kepler 5.4 Tycho licity are connected by lines: 5.4 Z (thick solid), 1.8 Z (thick dashed), 0.72 Z (thin solid), and 0.01 0.03 0.18 Z (thin dashed), The values for Kepler and Tycho indicated by the green and orange dia- 1.8 Z mondsMn/Fe are calculated using the line fluxes from Park et al. (2013) and Yamaguchi et al. (2014). Z (b) Same as panel0.72 (a), but the values predicted for the innermost 0.2 M that is dominated by the n-NSE regime (blue) and the other regions (red) from the MCh models are separately shown. Both MCh DDT models The observed mass ratios for Blue: 3C n-NSE 397 region can be well explained by the that occurs in Z the n-NSE regions0.18 of standard Red: metallicityNSE + Si burning models. (Figures reproduced with the permission of the authors;0.001 0.003 Yamaguchi0.010.03 et al., 2015, 0.1 Figure 0.3 4, pp. 5). Ni/Fe

A key prediction of the SD scenario is the presence of circumstellar material (CSM). One possible source of this CSM are the fast, optically thick winds that are required to stabilize accretion and allow the WD to grow to MCh and avoid a common-envelope phase (Hachisu et al., 1996). These fast winds have mass loss −7 −1 rates comparable to the accretion rate (1-5 × 10 M yr ), and velocities larger than the escape velocity from the surface of a massive WD (several hundred km s −1), resulting in large mechanical . By the time of the SN explosion, these outflows should have carved out large (several pc in diameter) low-density cavities around the progenitor system. Even if accretion winds are not active, it is hard to envision any realization of the SD sce- nario that does not lead to some kind of CSM, because the accretion process cannot be 100% efficient. For outflow velocities slower than those found in accretion winds, the CSM should stay close to the progenitor, and geometric dilution will lead to a −2 density profile of the form ρCSM ∼ Ar , where A is a constant of proportionality Supernova Remnants as Clues to Their Progenitors 7 dependent upon the mass-loss rate and wind velocity. Badenes et al. (2007) studied the impact of several CSM distributions on SNR evolution, and concluded that most Type Ia SNRs show no significant evidence for the presence of CSM. In Figure 3 we show an updated comparison between the models of Badenes et al. (2007) and the Fe Kα centroids measured for Type Ia SNRs by Yamaguchi et al. (2014) (see Section 3 below for a more detailed explanation of these measure- ments). The physical radii and the ionization state of Fe in Type Ia SNRs are in good agreement with SNR models where Type Ia SN ejecta interact with an undisturbed −2 ambient medium (shaded regions). Models with a ρCSM ∼ Ar circumstellar pro- file (blue plot) lead to small, highly ionized SNRs that do not appear in the observed sample (though many core-collapse SNRs fit this description, see Section 3). Mod- els expanding into low density cavities excavated by accretion winds lead to large, low-ionization SNRs. Although these objects are not common, there is at least one example in the sample from Yamaguchi et al. (2014): SNR RCW 86. This SNR has long been suspected to be a cavity explosion, and its bulk dynamics are hard to in- terpret without a fast, sustained outflow from the progenitor, with properties similar to those of accretion winds (Badenes et al., 2007; Williams et al., 2011).

Fig. 3 Fe Kα line centroid as a function of radius for the Type Ia SNRs in Yamaguchi et al. (2014), plus DEM L71 (Maggi et al., 2016). The shaded regions represent theoretical predictions from Type Ia SNR models interacting with a uniform interstellar medium (Badenes et al., 2003, 2005): orange for the 56Ni-poor delayed detonation model DDTg, blue for bright 56Ni-rich delayed detonation model DDTa. The colored lines represent theoretical predictions from Badenes et al. (2007) for model DDTg interacting with two circumstellar profiles calculated assuming an SD progenitor, red for an accretion wind outflow with a large cavity (SDfast) and blue for a ρ ∝ r−2 outflow (SDslow).

A curious feature of Figure 3 is the agreement shown between Kepler’s SNR and models for an unmodified ambient medium. Shown in Figure 4 (left) is an X-ray image of Kepler, showing clear signs of circumstellar interaction. Patnaude et al. (2012) modeled the dynamics and X-ray emission for a Type Ia model expanding in the dense wind of an AGB companion (Chiotellis et al., 2012). Patnaude et al. (2012) jointly modeled the blastwave dynamics and X-ray spectrum of the SNR. The line 8 Daniel Patnaude and Carles Badenes energy centroids of prominent emission lines of silicon and are sensitive to the circumstellar environment the remnant is expanding into. Patnaude et al. (2012) found that in order to match both the dynamics and bulk X-ray spectral properties, a small cavity, of order 0.1 pc in radius, is required around the progenitor. They suggest that the cavity could have been formed just prior to thermonuclear runaway, as the accretion wind pushes out against the dense wind from the AGB companion. Given the high Fe content observed in the spectrum of Kepler’s SNR (Patnaude et al., 2012; Katsuda et al., 2015a), there is little doubt that it is the result of a Ia and the CSM wind scenario argued by Chiotellis et al. (2012); Patnaude et al. (2012); Katsuda et al. (2015a) seems to favor the SD channel, though fast accretion winds appear ruled out. However, Kerzendorf et al. (2014) obtained spectra for 24 stars with L > 10L , and none showed evidence for being a former donor star, including no , AGB, or post-AGB surviving companions in the inner 4000 of the rem- nant. Similar searches for companions in the LMC remnants SNR 0519-67.5 and 0509-67.5 also turn up no evidence for a surviving companion (Schaefer & Pag- notta, 2012; Edwards et al., 2012). While this isn’t firm proof for the viability of the DD scenario, it does suggest that, at least in these cases, the double degenerate progenitor model could be required in order to produce these SNRs. To conclude, observations of Type Ia SNRs provide clear evidence for a MCh SN Ia progenitor in SNR 3C 397 (Yamaguchi et al., 2015), and the presence of a fast, sustained outflow in SNR RCW 86 (Badenes et al., 2007; Williams et al., 2011). Taken at face value, this seems to support the SD scenario for at least some local Type Ia SNe. However, the bulk properties of most Type Ia SNRs are at odds with the CSM configurations expected in SD progenitors (see Figure 3 and Badenes et al., 2007, for a detailed discussion). More contrived CSM profiles are certainly possible in some cases, as shown by Patnaude et al. (2012) for Kepler, but a more likely explanation is that DD progenitors also make a significant contribution to the local SN Ia rate (see Badenes & Maoz, 2012). A mixture of progenitor channels has interesting implications for cosmology (Ponder et al., 2016).

2.2 Core Collapse SNRs – Progenitor Connections

A good review concerning the connections between young supernova remnants and their progenitors can be found in Chevalier (2005). They note that different super- nova types arise form different zero age masses, with SNe IIP oc- curing in the 8-15 M range, SNe Ib/c occuring in stars with ZAMS mass & 35 M , and IIb/IIL SNe resulting from stars in the middle, with observational results providing information on the progenitor to supernova type relationship Smartt (e.g., 2015). If firm connections between progenitor and supernova type can be established, then connecting a remnant to it’s parent supernova provides an indirect path back to the progenitor type. Chevalier (2005) attempted this for several objects, including the Crab, 3C 58, and G292.0+1.8. In the Crab, measured abundances suggest that the Supernova Remnants as Clues to Their Progenitors 9

30"

Fig. 4 Left: Kepler’s SNR viewed in X-rays with Chandra ACIS-S3. The RGB image shows 0.4– 0.75 keV emission in red, 0.75–1.2 keV emission in green, and 1.2–7.0 keV emission in blue. In the image, north is up and east is to the left. The outlined region corresponds to the blastwave–ejecta interaction discussed in Patnaude et al. (2012). Right: In the upper panel, we plot the measured versus computed line centroids for Si Kα,SKα, and Fe Kα, for the DDTa (black solid) and DDTg (red dashed) models. The hatched region in each panel corresponds to the measured centroid including the 90% confidence interval The DDTg models and a subset of the DDTa models (M˙ < −6 −1 4×10 M yr ) do not produce any Fe K emission. In the lower panel, we plot the line centroids −1 for the DDTa cavity models for vwind = 10 (green solid) and 20 (blue dashed) km s , for a range of mass-loss rates. The line centroids and errors are indicated by the vertical hatched regions in each panel. The allowed mass-loss rates as dictated by the comparison between the measured and modeled line centroids are marked by the horizontal cross hatched region (Figures reproduced with the permission of the authors; Patnaude et al., 2012). progenitor lacked an O-rich mantle, placing the progenitor star at the low end of the supernova progenitor mass range, ∼ 8–10 M , suggesting that the Crab supernova was of the Type IIP variety. On the other hand, the SNR G292.0+1.8 probably arises from a more massive progenitor, as the blastwave has swept up ∼ 15–40 M of RSG wind (Lee et al., 2010). Estimates of a high swept up mass points to a massive progenitor, with an initial mass of 20–35 M . In general, the range of progenitor models which lead to core collapse super- nova remnants is much larger than one sees in the case of Type Ia SNRs. This is reflected in the morphological and compositional diversity of CC SNR we observe in the . Ranges in mass loss rates and wind speeds, combined with the initial mass of the star (which will influence the final abundances), as well as evolution- ary changes that the progenitor may go through (such as a Wolf-Rayet phase), not to mention the effects that a binary companion can have on the progenitor’s evolu- tion, leads to larger parameter space to explore when comparing CC SNR to their progenitors. As seen in Lee et al. (2010) and others, examples exist where broad statements about the progenitor can be made, but few objects allow for the same types of detailed analyses that we find in Ia studies. The exceptions are the Galac- 10 Daniel Patnaude and Carles Badenes tic SNR (Cas A), and the youngest know remnant, SN 1987A, in the Large Magallenic Cloud. Cas A, shown in Figure 5 (left), may represent the most well studied of the Galac- tic SNR. At an age of 330 yr (Thorstensen et al., 2001), it has a morphology which is consistent with that of a SNR interacting with a RSG wind (Chevalier & Oishi, 2003). Lee et al. (2014); Hwang & Laming (2012) undertook detailed studies of the mass-loss history of the Cas A progenitor, and estimate that the progenitor star shed & 6M of material prior to core–collapse. Combined with the estimated mass of the neutron star and ejecta, the progenitor zero-age main sequence mass is ≈ 12 M .

Fig. 5 Left: Cas A in X-rays. The three color image shows 0.5–1.5 keV in red, 1.5–3.0 keV in green, and 4.2–6.0 keV in green. The neutron star is clearly visible near the center. Right: Density of the supernova remnant at a time of 330 yr after explosion. The left panel shows the remnant that results from the evolution of ejecta in a CSM modified by a pure RSG wind, with the middle and righthand panels showing the case where the Wolf-Rayet phase lasted for 2265 and 3480 years, respectively. The forward shock of the remnant in all three cases is located at a distance of ∼ 2.4 pc, while the reverse shock is located increasingly farther inward for longer WR phases Details of the models may be found in Schure et al. (2008). (Righthand figure reproduced with the permission of the authors; Schure et al., 2008, Figure 4, pp 403).

The spectra of the Cas A SN obtained by Krause et al. (2008); Rest et al. (2011) suggest a Type IIb origin, similar to SN 1993J, implying a red supergiant origin. The analysis by Lee et al. (2014) suggest that the Cas A progenitor lost much of its mass through the RSG wind, compatible with the IIb SN type, though this is at odds with the mass-loss rates adopted for single RSG stars (Woosley et al., 2002). Solutions to the high amount of mass lost from Cas A include episodic mass loss through pulsational instabilities in the RSG phase Yoon & Cantiello (2010), or enhanced mass loss through the interaction of a binary companion (Young et al., 2006). Another possibility is that the Cas A progenitor went through a short Wolf-Rayet phase prior to core collapse. The fast wind from a Wolf-Rayet progenitor would clear out a small cavity, much like is required in Kepler’s SNR, and would imprint Supernova Remnants as Clues to Their Progenitors 11 itself on the dynamics and X-ray emission of the remnant. Hwang & Laming (2012) required such a cavity in order to match the observed X-ray emission. Schure et al. (2008) modeled the evolution of the Cas A jet in a cavity, and found that if a Wolf-Rayet phase existed, it would be rather short (∼ a few thousand years; Fig. 5, right). Besides Cas A, perhaps the most well studied supernova remnant is that of SN 1987A, a recent review of which can be found in McCray & Fransson (2016). SN 1987A is the closest example of a remnant where the progenitor is known through pre-explosion imaging. However, supernovae are now regularly identified with their progenitors through the use of pre-explosion images. Smartt (2015) notes 18 detections of supernova progenitors, with 27 additional upper limits – a large number of the sample of detected supernova progenitors result in SN IIP, IIL, or IIb. This is consistent with Chevalier (2005), who associated many Galactic SNR with Type II SNe, resulting from RSG progenitors of various masses. Only one or possibly two Galactic SNR have been firmly associated with SNe Type Ib/c (and hence with a Wolf-Rayet progenitor) – W49B as a possible gamma-ray burst rem- nant (Lopez et al., 2013), and RX J1713–3946 (Katsuda et al., 2015b).

3 SNR Bulk Properties

As discussed in the previous sections, typing individual supernova remnants remains challenging. Recently. Yamaguchi et al. (2014) presented a method of typing SNRs based on the Fe-Kα line centroid and . Since Fe is produced in the center of the progenitor during the explosion, heating of Fe can be delayed, resulting in ionization states lower than He-like (Fe24+) in young and middle-aged SNRs. The ionization state affects the Fe-K line centroid, which can be measured with high precision with satellites such as Chandra, XMM–Newton, or Suzaku. In brief, they found a correlation between supernova type, and the Fe-K line centroid: the line centroid for Type Ia SNRs are generally lower (< 6550 eV) than those found in CCSNRs. Perhaps more importantly, when they computed synthetic line centroids from models for Type Ia ejecta interacting with a uniform medium, the models pre- dict bulk properties in line with the observations. Patnaude et al. (2015) extended this work to study core collapse SNR progen- itors. While Yamaguchi et al. (2014) focussed on Type Ia SNR in an unmodified circumstellar environment, a key question, as demonstrated in previous sections, is the nature of the circumstellar environment around massive progenitors. Red and Yellow supergiants (R/YSGs) expel several solar masses of material over their life- −5 −1 times, with mass loss rates ≈ 10 M yr . Wolf-Rayet stars, thought to be the progenitors of the Ib/c subset of core collapse supernovae, expel mass at similar rates, but with wind velocities two orders of higher. In addition to steady line-driven mass loss, episodic mass loss may also be rele- vant. Patnaude et al. (2015) modeled the circumstellar environments of core collapse progenitors for a range of mass loss rates and wind velocities, but did not consider 12 Daniel Patnaude and Carles Badenes

W49B N103B Cas A N132D 0519-69.0 Cas A W49B N132D N63A 3C 397 N63A Sgr A East Sgr A East Kepler SN 1987A G349.7+0.2 N49 0509-67.5 G344.7-0.1 N49 SN 1987A G0.61+0.01 Tycho G292.0+1.8 G349.7+0.2 G350.1-0.3 G0.61+0.01 IC 443

RCW86 G292.0+1.8 G352.7-0.1 G350.1+0.3

SN 1006 G337.2-0.7 IC 443

Fig. 6 Left: Fe-K line luminosity vs. centroid energy for Galactic and Magallenic Cloud Ia and core collapse supernova remnants measured by Yamaguchi et al. (2014). The transparent shaded regions correspond to models for either Ia or CCSNe ejecta, with yellow and light blue corresponding to the Ia models DDTa and DDTg, and red, green, magenta, and dark blue corresponding to CCSNe models SN 1987A, SN 1993J, and a 12 and 25M solar metallicity progenitor, respectively. The data are from Yamaguchi et al. (2014). Right: Fe-K line luminosity as a function of radius for core collapse supernova remnant measurements and models (Figures reproduced with the permission of the authors; Patnaude et al., 2015). the effects that episodic mass loss would have on their results. In Figure 6 (left), we reproduce the results from Patnaude et al. (2015). As seen in the left hand panel, there is a clear bifurcation between Type Ia and CCSNe ejecta models, consistent with the observed properties of the remnants. In particular, the CCSNe models con- sistently produce higher Fe-K line centroids. Patnaude et al. (2015) attribute this to the blastwave interacting with a great deal of material deposited by the stellar wind around the progenitor. While these models capture the general spectral characteristics of the data, when correlating modeled radii to line luminosity, an apparent disconnect is seen between the modeled line luminosities as a function of radius, as compared to the data. This is shown in Figure 6 (right), where the models underestimate the Fe-K line luminosity compared to the data. In the case of CCSNe, the Fe-K line centroid is already close to the collisional ionization equilibrium value. Modest increases in the progenitor mass loss rate will have a small effect on the ionization, while providing a boost to the line luminosity, since the X-ray emission scales like the square of the density. Increases in the mass loss rate, if non-steady, would not affect the radius, as the interaction timescale between the blast wave and the additional circumstellar material would be short, compared to the overall age of the remnant. There is now growing evidence that core collapse progenitors undergo some sort of episodic mass loss prior to core collapse. The optical light curves of several Type IIn supernovae, including SN 2005ip, SN 2005kj, and SN 2006aa, all suggest Supernova Remnants as Clues to Their Progenitors 13 episodic mass loss prior to core collapse (Moriya et al., 2014), and a mass loss event from the progenitor of SN 2009ip was sufficiently luminous that it was misclassified as a SN, though the progenitor did finally explode in 2012. The mechanism for the episodic mass loss is not currently understood, but the mass ejection may be related to wave-driven mass loss during core and oxygen burning (Shiode & Quataert, 2014), or possibly unsteady nuclear burning in the envelope (Smith, 2014). The mass loss history, steady or otherwise, will imprint itself on the X-ray emis- sion from the SNR (Patnaude et al., 2015). As an example of this, Moriya (2012) investigated the progenitors for so-called recombining supernova remnants, rem- nants where the ion temperature is sufficiently higher than the electron temperature, and found that a plausible explanation exists for them if their progenitors lost a suffi- cient amount of mass that was subsequently deposited near the progenitor, just prior to core-collapse. They argued that the progenitors of these remnants are likely red supergiants, and are likely not associated with SNe IIn.

4 Conclusions

In this chapter we have reviewed several properties of supernova remnants which provide links back to their progenitor systems. X-ray observations of supernova remnants place constraints on the metallicity of the progenitor and measuring the explosion energetics, either through proper motion analyses or inferred from spec- tral measurements can provide constraints on the explosion energetics, as well as the density of the circumstellar medium. These parameters relate directly back to the evolution of the progenitor. In the case of Type Ia SNRs, measuring the abundances of neutron-rich material places constraints on the central density of the white dwarf progenitor. Inferring a high central density points to a MCh explosion and likely a single-degenerate pro- genitor system. Likewise, joint spectral/dynamical fits to objects such as Kepler’s SNR point to complicated evolutionary histories for the progenitor, where an AGB wind is required from the progenitor system, but only if it has a cavity excised out of it prior to the supernova event. A similar result has been suggested in Tycho (Chiotellis et al., 2013; Slane et al., 2014). For progenitors od core-collapse SNR, the situation is more complicated. CCSNe progenitors exhibit a larger mass range, with order of magnitude variations in mass loss rates and progenitor wind velocities. Except for the notable cases of SN 1987A and Cas A, only broad statements about the SNR progenitor may be made at this time. However, when one begins to step back from individual remnants, trends in SNR bulk properties begin to emerge. Recent results by Yamaguchi et al. (2014) show a clear bifurcation in line centroids between SNR Ia and core-collapse SNRs. These results are consistent with model predictions. While models for SNR progenitors do generally agree with observed bulk prop- erties of SNR, there are still inconsistencies. Nonetheless, these issues may be ad- dressed by stellar evolution models which better define the end stages of the pro- 14 Daniel Patnaude and Carles Badenes genitor, including such affects as episodic mass loss and non-steady nuclear burn- ing. Both mechanisms alter the circumstellar environment in a way that is readily observable in remnants, hundreds of years after the supernova event.

Acknowledgments

The authors wish to thank the editorial staff for their steadfast patience during the preparation of this manuscript. D. J. P. acknowledges support from NASA contract NAS8-03060 and NASA/TM6-17003X. C. B. acknowledges partial support from grants NSF/AST-1412980 and NASA/NNX15AM03G-S01.

Cross References

• Supernova Remnant Cassiopeia A • Observational Classification of Supernovae • Dynamical Evolution and Radiative Processes in Supernova Remnants • Supernova remnant from SN 1987A

References

Badenes, C., Bravo, E., Borkowski, K. J., & Dom´ınguez, I. 2003, “Thermal X-Ray Emission from Shocked Ejecta in Remnants: Prospects for Explosion Mechanism Identification,” ApJ, 593, 358 Badenes, C., Borkowski, K. J., & Bravo, E. 2005, “Thermal X-Ray Emission from Shocked Ejecta in Type Ia Supernova Remnants. II. Parameters Affecting the Spectrum,” ApJ, 624, 198 Badenes, C., Borkowski, K. J., Hughes, J. P., Hwang, U., & Bravo, E. 2006, “Con- straints on the Physics of Type Ia Supernovae from the X-Ray Spectrum of the Tycho Supernova Remnant,” ApJ, 645, 1373 Badenes, C., Hughes, J. P., Bravo, E., & Langer, N. 2007, “Are the Models for Type Ia Supernova Progenitors Consistent with the Properties of Supernova Rem- nants?,” ApJ, 662, 472 Badenes, C., Harris, J., Zaritsky, D., & Prieto, J. L. 2009, “The Stellar Ancestry of Supernovae in the Magellanic Clouds. I. The Most Recent Supernovae in the ,” ApJ, 700, 727 Badenes, C. 2010, “X-ray studies of supernova remnants: A different view of super- nova explosions,” Proceedings of the National Academy of , 107, 7141 Badenes, C., & Maoz, D. 2012, “The Merger Rate of Binary White Dwarfs in the Galactic Disk,” ApJ, 749, L11 Supernova Remnants as Clues to Their Progenitors 15

Borkowski, K. J., Lyerly, W. J., & Reynolds, S. P. 2001, “Supernova Remnants in the Sedov Expansion Phase: Thermal X-Ray Emission,” ApJ, 548, 820 Borkowski, K. J., Reynolds, S. P., Green, D. A., et al. 2014, “Nonuniform Expansion of the Youngest Galactic Supernova Remnant G1.9+0.3,” ApJ, 790, L18 Brachwitz, F., Dean, D. J., Hix, W. R., et al. 2000, “The Role of Electron Captures in Chandrasekhar-Mass Models for Type IA Supernovae,” ApJ, 536, 934 Chevalier, R. A., & Oishi, J. 2003, “Cassiopeia A and Its Clumpy Presupernova Wind,” ApJ, 593, L23 Chevalier, R. A. 2005, “Young Core-Collapse Supernova Remnants and Their Su- pernovae,” ApJ, 619, 839 Chiotellis, A., Schure, K. M., & Vink, J. 2012, “The imprint of a symbiotic binary progenitor on the properties of Kepler’s supernova remnant,” A&A, 537, A139 Chiotellis, A., Kosenko, D., Schure, K. M., Vink, J., & Kaastra, J. S. 2013, “Mod- elling the interaction of thermonuclear supernova remnants with circumstellar structures: the case of Tycho’s supernova remnant,” MNRAS, 435, 1659 Chomiuk, L., Soderberg, A. M., Chevalier, R. A., et al. 2016, “A Deep Search for Prompt Radio Emission from Thermonuclear Supernovae with the Very Large Array,” ApJ, 821, 119 DeLaney, T., & Rudnick, L. 2003, “The First Measurement of Cassiopeia A’s For- ward Shock Expansion Rate,” ApJ, 589, 818 Dessart, L., Blondin, S., Hillier, D. J., & Khokhlov, A. 2014, “Constraints on the explosion mechanism and progenitors of Type Ia supernovae,” MNRAS, 441, 532 Edwards, Z. I., Pagnotta, A., & Schaefer, B. E. 2012, “The Progenitor of the Type Ia Supernova that Created SNR 0519-69.0 in the Large Magellanic Cloud,” ApJ, 747, L19 Ferriere,` K. M. 2001, “The interstellar environment of our galaxy,” Reviews of Mod- ern Physics, 73, 1031 Filippenko, A. V. 1997, “Optical Spectra of Supernovae,” ARA&A, 35, 309 Hachisu, I., Kato, M., & Nomoto, K. 1996, “A New Model for Progenitor Systems of Type IA Supernovae,” ApJ, 470, L97 Hillebrandt, W., Kromer, M., Ropke,¨ F. K., & Ruiter, A. J. 2013, “Towards an un- derstanding of Type Ia supernovae from a synthesis of theory and observations,” Frontiers of Physics, 8, 116 Hwang, U., & Laming, J. M. 2009, “The Circumstellar Medium of Cassiopeia a Inferred from the Outer Ejecta Knot Properties,” ApJ, 703, 883 Hwang, U., & Laming, J. M. 2012, “A Chandra X-Ray Survey of Ejecta in the Cassiopeia A Supernova Remnant,” ApJ, 746, 130 Jennings, Z. G., Williams, B. F., Murphy, J. W., et al. 2014, “The Supernova Progen- itor Mass Distributions of M31 and M33: Further Evidence for an Upper Mass Limit,” ApJ, 795, 170 Katsuda, S., Mori, K., Maeda, K., et al. 2015, “Kepler’s Supernova: An Overlumi- nous Type Ia Event Interacting with a Massive Circumstellar Medium at a Very Late Phase,” ApJ, 808, 49 Katsuda, S., Acero, F., Tominaga, N., et al. 2015, ApJ, 814, 29 16 Daniel Patnaude and Carles Badenes

Kerzendorf, W. E., Childress, M., Scharwachter,¨ J., Do, T., & Schmidt, B. P. 2014, “A Reconnaissance of the Possible Donor Stars to the Kepler Supernova,” ApJ, 782, 27 Koo, B.-C., & McKee, C. F. 1992, “Dynamics of wind bubbles and . I - Slow winds and fast winds. II - Analytic theory,” ApJ, 388, 93 Krause, O., Birkmann, S. M., Usuda, T., et al. 2008, “The Cassiopeia A Supernova Was of Type IIb,” Science, 320, 1195 bibitem[Kerzendorf et al.(2014)]kerzendorf14 Kerzendorf, W. E., Childress, M., Scharwachter,¨ J., Do, T., & Schmidt, B. P. 2014, “A Reconnaissance of the Pos- sible Donor Stars to the Kepler Supernova,” ApJ, 782, 27 Lee, J.-J., Park, S., Hughes, J. P., et al. 2010, “The Outer Shock of the Oxygen-Rich Supernova Remnant G292.0+1.8: Evidence for the Interaction with the Stellar Winds from Its Massive Progenitor,” ApJ, 711, 861 Lee, J.-J., Park, S., Hughes, J. P., & Slane, P. O. 2014, “X-Ray Observation of the Shocked Red Supergiant Wind of Cassiopeia A,” ApJ, 789, 7 Lopez, L. A., Ramirez-Ruiz, E., Castro, D., & Pearson, S. 2013, “The Galactic Su- pernova Remnant W49B Likely Originates from a Jet-driven, Core-collapse Ex- plosion,” ApJ, 764, 50 Maggi, P., Haberl, F., Kavanagh, P. J., et al. 2016, “The population of X-ray super- nova remnants in the Large Magellanic Cloud,” A&A, 585, A162 Maoz, D., Mannucci, F., & Nelemans, G. 2014, “Observational Clues to the Progen- itors of Type Ia Supernovae,” ARA&A, 52, 107 Margutti, R., Parrent, J., Kamble, A., et al. 2014, “No X-Rays from the Very Nearby Type Ia SN 2014J: Constraints on Its Environment,” ApJ, 790, 52 Maund, J. R., Smartt, S. J., Kudritzki, R. P., Podsiadlowski, P., & Gilmore, G. F. 2004, “The massive binary companion star to the progenitor of supernova 1993J,” Nature, 427, 129 McCray, R., & Fransson, C. 2016, “The Remnant of Supernova 1987A,” ARA&A, 54, 19 Moriya, T. J. 2012, “Progenitors of Recombining Supernova Remnants,” ApJ, 750, L13 Moriya, T. J., Maeda, K., Taddia, F., et al. 2014, “Mass-loss histories of Type IIn su- pernova progenitors within decades before their explosion,” MNRAS, 439, 2917 Park, S., Badenes, C., Mori, K., et al. 2013, “A Super-solar Metallicity for the Pro- genitor of Kepler’s Supernova,” ApJ, 767, L10 Patnaude, D. J., & Fesen, R. A. 2009, “Proper Motions and Brightness Variations of Nonthermal X-ray Filaments in the Cassiopeia A Supernova Remnant”, ApJ, 697, 535 Patnaude, D. J., Badenes, C., Park, S., & Laming, J. M. 2012, “The Origin of Ke- pler’s Supernova Remnant,” ApJ, 756, 6 Patnaude, D. J., Lee, S.-H., Slane, P. O., et al. 2015, “Are Models for Core-collapse Supernova Progenitors Consistent with the Properties of Supernova Remnants?,” ApJ, 803, 101 Supernova Remnants as Clues to Their Progenitors 17

Perez-Torres,´ M. A., Lundqvist, P., Beswick, R. J., et al. 2014, “Constraints on the Progenitor System and the Environs of SN 2014J from Deep Radio Observa- tions,” ApJ, 792, 38 Ponder, K. A., Wood-Vasey, W. M., & Zentner, A. R. 2016, “Incorporating Astro- physical Systematics into a Generalized Likelihood for Cosmology with Type Ia Supernovae,” ApJ, 825, 35 Raymond, J. C., Korreck, K. E., Sedlacek, Q. C., et al. 2007, “The Preshock Gas of SN 1006 from Advanced Camera for Surveys Observa- tions,” ApJ, 659, 1257 Rest, A., Welch, D. L., Suntzeff, N. B., et al. 2008, “Scattered-Light Echoes from the Historical Galactic Supernovae Cassiopeia A and Tycho (SN 1572),” ApJ, 681, L81 Rest, A., Foley, R. J., Sinnott, B., et al. 2011, “Direct Confirmation of the Asymme- try of the Cas A Supernova with Light Echoes,” ApJ, 732, 3 Saio, H., & Nomoto, K. 1985, “Evolution of a merging pair of C+O white dwarfs to form a single neutron star,” A&A, 150, L21 Schaefer, B. E., & Pagnotta, A. 2012, “An absence of ex-companion stars in the type Ia supernova remnant SNR 0509-67.5,” Nature, 481, 164 Schure, K. M., Vink, J., Garc´ıa-Segura, G., & Achterberg, A. 2008, “Jets as Diag- nostics of the Circumstellar Medium and the Explosion Energetics of Supernovae: The Case of Cassiopeia A,” ApJ, 686, 399-407 Shen, K. J., Guillochon, J., & Foley, R. J. 2013, “Circumstellar Absorption in Dou- ble Detonation Type Ia Supernovae,” ApJ, 770, L35 Shen, K. J., & Bildsten, L. 2007, “Thermally Stable Nuclear Burning on Accreting White Dwarfs,” ApJ, 660, 1444 Shiode, J. H., & Quataert, E. 2014, “Setting the Stage for Circumstellar Interaction in Core-Collapse Supernovae. II. Wave-driven Mass Loss in Supernova Progeni- tors,” ApJ, 780, 96 Slane, P., Lee, S.-H., Ellison, D. C., et al. 2014, “A CR-hydro-NEI Model of the Structure and Broadband Emission from Tycho’s Supernova Remnant ,” ApJ, 783, 33 Smartt, S. J., Maund, J. R., Gilmore, G. F., et al. 2003, “Mass limits for the progen- itor star of supernova 2001du and other Type II-P supernovae,” MNRAS, 343, 735 Smartt, S. J. 2015, “Observational Constraints on the Progenitors of Core-Collapse Supernovae: The Case for Missing High-Mass Stars,” Publications of the Astro- nomical Society of Australia, 32, e016 Smith, N. 2014, “Mass Loss: Its Effect on the Evolution and Fate of High-Mass Stars,” ARA&A, 52, 487 Thorstensen, J. R., Fesen, R. A., & van den Bergh, S. 2001, “The Expansion Center and Dynamical Age of the Galactic Supernova Remnant Cassiopeia A,” AJ, 122, 297 van Kerkwijk, M. H., Chang, P., & Justham, S. 2010, “Sub-Chandrasekhar White Dwarf Mergers as the Progenitors of Type Ia Supernovae,” ApJ, 722, L157 18 Daniel Patnaude and Carles Badenes

Van Dyk, S. D., Garnavich, P. M., Filippenko, A. V., et al. 2002, “The Progenitor of Supernova 1993J Revisited,” PASP, 114, 1322 Van Dyk, S. D., Li, W., & Filippenko, A. V. 2003, “On the Progenitor of the Type II-Plateau Supernova 2003gd in M74,” PASP, 115, 1289 Vink, J. 2012, “Supernova remnants: the X-ray perspective,” A&A Rev., 20, 49 Wang, B., & Han, Z. 2012, “Progenitors of type Ia supernovae,” New A Rev., 56, 122 Williams, B. J., Blair, W. P., Blondin, J. M., et al. 2011, “RCW 86: A Type Ia Su- pernova in a Wind-blown Bubble,” ApJ, 741, 96 Williams, B. F., Peterson, S., Murphy, J., et al. 2014, “Constraints for the Progenitor Masses of 17 Historic Core-collapse Supernovae,” ApJ, 791, 105 Woosley, S. E., Heger, A., & Weaver, T. A. 2002, “The evolution and explosion of massive stars,” Reviews of Modern Physics, 74, 1015 Yamaguchi, H., Badenes, C., Petre, R., et al. 2014, “Discriminating the Progenitor Type of Supernova Remnants with Iron K-shell Emission,” ApJ, 785, L27 Yamaguchi, H., Badenes, C., Foster, A. R., et al. 2015, “A Chandrasekhar Mass Progenitor for the Type Ia Supernova Remnant 3C 397 from the Enhanced Abun- dances of and Manganese,” ApJ, 801, L31 Yang, X. J., Tsunemi, H., Lu, F. J., et al. 2013, “Cr-K Emission Line as a Constraint on the Progenitor Properties of Supernova Remnants,” ApJ, 766, 44 Yoon, S.-C., & Cantiello, M. 2010, “Evolution of Massive Stars with Pulsation- driven Superwinds During the Red Supergiant Phase,” ApJ, 717, L62 Young, P. A., Fryer, C. L., Hungerford, A., et al. 2006, “Constraints on the Progeni- tor of Cassiopeia A,” ApJ, 640, 891 Zhou, P., Chen, Y., Zhang, Z.-Y., et al. 2016, “Expanding Molecular Bubble Sur- rounding Tycho’s Supernova Remnant (SN 1572) Observed with the IRAM 30m Telescope: Evidence for a Single-degenerate Progenitor,” ApJ, 826, 34